← Back to community benchmarks

Qwen3.5-9B

M3 Ultra (60c) · 96 GB · 4bit · 2026-05-18

Performance

32k

tokens

1,014

PP tok/s

84.5

TG tok/s

32326

TTFT (ms)

9.7

Peak mem (GB)

Hardware

Chip M3 Ultra (60c)

Memory 96 GB

GPU Cores 60

Software

oMLX v0.3.8

macOS macOS 26.4.1

Context 32,768

Performance by Context Length

Context	PP tok/s	TG tok/s	Peak Mem
1k	907.6	110.8	5.8 GB	view
4k	1,074	108.4	6.5 GB	view
8k	1,108	106.5	6.9 GB	view
16k	1,077	97.7	7.9 GB	view
32k	1,014	84.5	9.7 GB	current
64k	898.2	66.9	13.2 GB	view
128k	692.1	46.5	20.6 GB	view
195k	555.3	34.9	29.2 GB	view